Settlement Agreements under French Sapin II Law: In Search of the ‘Public Interest’

by Luca d’Ambrosio

This post is an abstract of the article forthcoming in the Revue de sciences criminelles et droit comparé (n° 1/2019) under the title L’implication des acteurs privés dans la lutte contre la corruption: un bilan en demi-tente de la loi Sapin 2.


Much has been reported about the adoption, on December 2016, of the new French anticorruption framework, Sapin II, which  stands out for the creation of a new set of ex ante and ex post measures aiming to strengthen the prevention of corruption and the enforcement of administrative and criminal sanctions.

Among the ex post measures, Sapin II introduced a procedure permitting a negotiated outcome for legal persons: under the name of “convention judiciaire d’intérêt public” (CJIP), this procedure emulates DPAs as practiced in the United States and in the United Kingdom. The legal transplant of this procedure into the French enforcement system has received far from unanimous consent.   

On the one hand, French scholars were divided among those who considered this procedure as a “gift” to corporations and those who considered it as a milestone of a new and effective corporate enforcement policy based on compliance and cooperation. According to this view, settlement agreements would enhance corporate enforcement policy for three reasons. Firstly, they would help enforcement authorities to resolve quickly and costless complex criminal cases. Secondly, they would enhance specific deterrence of future misconduct through remedial compliance programs. Finally, settlement agreements would trigger anticorruption cooperation and enforcement with US authorities: this argument was particularly sensible in France where important and strategic companies – such as Alstom, Société Générale, Total et Alcatel – have been involved in FCPA investigations and are in the “top ten” of the most important fines settled by the DOJ.

On the other hand, several institutions argued that the introduction of settlement agreements could threaten the legitimacy of criminal enforcement policy. In its opinion (PDF: 232 KB) on the Sapin II bill, the Conseil d’Etat considered the CJIP inconsistent with the mission of criminal justice, i.e. to guarantee exemplarity and truth. Another critical opinion (PDF: 417 KB) was delivered by the National Advisory Commission on Human Rights who stressed that CJIP may be inconsistent with the rule of law: despite the last evolutions, French public prosecutors are in fact not completely independent from the government and that Sapin II does not grant enough power to the judges to restrict the process governing prosecutors’ use of CJIP. As Jennifer Arlen stressed about the practice of using DPAs in the US, the empowerment of prosecutors without a genuine external oversight on the DPAs procedure may in fact “divert criminal enforcement to achieve idiosyncratic conception of public interest.”


Faced with this risk, it remains to be seen when the employment of a CJIP may be considered, as opposed to prosecution, in the public interest.

Beyond the semantic choice – “public interest” is in fact inscribed with some emphasis in both the agreement and the fine that companies accept to pay – the French legislature gives no elements on this point. It fixes in fact the terms of the agreement (a fine based on the revenues derived from the offence; the eventual reparation of damages to the victims of the offence; the set up a remediation compliance program for a maximum period of three years), but it is silent on the criteria for entering into the agreement.

In this aim, it is necessary to look at the Guidelines on the criminal matters of Sapin II published on January 2018, by the Ministry of Justice (PDF: 709 KB). This text identifies three criteria to proceed with a CJIP: (i) the antecedents of the company; (ii) voluntary self-report of the misconducts; (iii) cooperation with the enforcement authority. Whilst the first criteria will “in most cases lead to exclude the settlement agreement when the legal person has already benefited from such a measure,” the other two criteria will lead to a conclusion that a CJIP is more likely in the public interest than prosecution.

Despite the little space that the guidelines devote to such a fundamental issue (half page over 37…), this document permits to affirm that “French DPAs” do not differ from those envisaged in other countries. In the United States, self-reporting and cooperation are, among the ten criteria envisaged by the U.S. Attorney’s Manual, the most relevant criteria considered by prosecutors when deciding whether to employ DPAs or NPAs. In the United Kingdom, the proactive approach adopted by the management of the company when the offending is brought to their notice is considered as one of the main “interest of justice factors” against prosecution by the DPAs Code of Practice (PDF:346 KB). In Argentina, where a settlement agreement procedure has been recently introduced by the 2017 Anticorruption Act, the “effective cooperation” between the company and the prosecutor is one of the legal conditions listed to avoid prosecution by entering into the “Acuerdo de colaboración eficaz”.

Therefore, CJIP dramatically breaks with the vertical and inquisitorial view of justice which is still dominant in France. It also breaks with the general regime of corporate criminal liability which did not previously know any positive incentive to curb the asymmetry of information that exists between public and private actors about corporate crimes by inducing companies to help enforcement authorities to deter these crimes.  


It remains to be seen how the achievement of this goal may be guaranteed in practice. In this regard, the judicial control on the settlement agreement procedure is crucial. The oversight of an independent third-party not only limits the prosecutors’ discretion in the settlement agreement procedure. It also permits to transcend the particular or idiosyncratic interests that necessarily come into play during the negotiation of a settlement into general interest. This is why “French DPAs” are submitted to judicial approval. This procedure involves (i) the regularity of the procedure and (ii) the proportionality of the fine and the other measures in relation to the benefits that the company received from the offence. It may be argued, especially after the publication of the guidelines, that the “control of the regularity of the procedure” involves also a substantial control on the criteria – self report and cooperation – that should lead the prosecutor to go through a settlement agreement rather than prosecution. However, the analysis of the first CJIP’s approval judgements suggests that French judges are rather comfortable in the role of “notary” that Sapin II allegedly entrusted.

In the first CJIP’s approval judgement issued on November 2017 in the HSBC case (PDF: 253 KB) relating to facts of tax fraud money laundering, the French judges limit to note that the company has “acknowledged the facts and explained the measures put in place to prevent further offences” and that “the prosecution justified in the public hearing the employment of the CJIP”. The same vague, and even more synthetic, formula has been used in a second CJIP’s approval judgment issued in June 2018 in the Société Générale case relating to international corruption of Libyan authorities (PDF: 93.3 KB). Despite the previous publication of the guidelines, this judgment is silent about cooperation and self-report of the bank. According to the judges, the issue has been systematically discussed in the public hearings. It is however highly regrettable that any track of this discussion is contained in the judgments. And this is even more regrettable as in the above cases the misconducts were respectively revealed by a whistleblower and by the press.

The first French CJIP’s approval judgements are thus strikingly different from those issued on the other side of the Channel by Sir Brian Levenson: in the approval judgments of the Standard Bank (2015), XYZ (2016) and Rolls-Royce (2017) DPAs concluded under the UK Bribery Act, different factors – such as the gravity of the offence, self-report, cooperation, legal compliance – are carefully and systematically evaluated and balanced by the judge to confirm that the employment of a DPA is “in the interest of justice”. This is particularly evident in the Rolls-Royce DPAs approval judgement: the cooperation of the company after the discovery of the misconducts was in fact so important for the judge that it permitted to counterbalance the gravity of the systematic and worldwide corruptive system organized by the company.


In conclusion, two main remarks can be drawn from this brief analysis of the criminal settlement procedure introduced in France.

Firstly, the main challenge for employing settlement agreements is to “weigh” the different factors that should be considered in the light (and not against) the public interest of a legitimate and effective corporate enforcement policy. The comparative analysis demonstrates that, among these factors, self-reporting and cooperation are the most relevant “public interest factors” to be considered when entering into a settlement agreement as an alternative to prosecution: in this perspective, negotiated settlements may in fact enhance, rather than weak, criminal enforcement policy.

Secondly, judicial oversight on settlement agreements is necessary to guarantee the achievement of this goal. But it may be insufficient if it does not involve the assessment of the above-mentioned factors. Of course, this “public interest test” may only be effective when the judges are provided a clear and binding standard to employ when determining whether entering into a settlement agreement is more likely in the public interest than prosecution.

Luca d’Ambrosio is Senior Research Fellow at Collège de France and lecturer at Sorbonne Law School. His research focuses on criminal, regulatory and civil enforcement, business ethics and compliance.


The views, opinions and positions expressed within all posts are those of the author alone and do not represent those of the Program on Corporate Compliance and Enforcement (PCCE) or of New York University School of Law.  PCCE makes no representations as to the accuracy, completeness and validity of any statements made on this site and will not be liable for any errors, omissions or representations. The copyright of this content belongs to the author and any liability with regards to infringement of intellectual property rights remains with the author.